What's New in Linux on IBM Z Systems

What's New in Linux on IBM z Systems

Martin Schwidefsky IBM Lab Böblingen, Germany 4th March 2015 Session 16450 IBM

Trademarks & Disclaimer The following are trademarks of the International Business Machines Corporation in the United States and/or other countries. For a complete list of IBM Trademarks, see www.ibm.com/legal/copytrade.shtml:

IBM, the IBM logo, BladeCenter, Calibrated Vectored Cooling, ClusterProven, Cool Blue, POWER, PowerExecutive, Predictive Failure Analysis, ServerProven, System p, System Storage, System x , z Systems, WebSphere, DB2 and Tivoli are trademarks of IBM Corporation in the United States and/or other countries. For a list of additional IBM trademarks, please see http://ibm.com/legal/copytrade.shtml.

The following are trademarks or registered trademarks of other companies: Java and all Java based trademarks and logos are trademarks of Sun Microsystems, Inc., in the United States and other countries or both Microsoft, Windows,Windows NT and the Windows logo are registered trademarks of Microsoft Corporation in the United States, other countries, or both. Intel, Intel logo, Intel Inside, Intel Inside logo, Intel Centrino, Intel Centrino logo, Celeron, Intel Xeon, Intel SpeedStep, Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. UNIX is a registered trademark of The Open Group in the United States and other countries or both. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. Cell Broadband Engine is a trademark of Sony Computer Entertainment Inc. InfiniBand is a trademark of the InfiniBand Trade Association. Other company, product, or service names may be trademarks or service marks of others.

NOTES: Linux penguin image courtesy of Larry Ewing ([email protected]) and The GIMP

Any performance data contained in this document was determined in a controlled environment. Actual results may vary significantly and are dependent on many factors including system hardware configuration and software design and configuration. Some measurements quoted in this document may have been made on development-level systems. There is no guarantee these measurements will be the same on generally-available systems. Users of this document should verify the applicable data for their specific environment. IBM hardware products are manufactured from new parts, or new and serviceable used parts. Regardless, our warranty terms apply. Information is provided “AS IS” without warranty of any kind. All customer examples cited or described in this presentation are presented as illustrations of the manner in which some customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics will vary depending on individual customer configurations and conditions.

This publication was produced in the United States. IBM may not offer the products, services or features discussed in this document in other countries, and the information may be subject to change without notice. Consult your local IBM business contact for information on the product or services available in your area. All statements regarding IBM's future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only. Information about non-IBM products is obtained from the manufacturers of those products or their published announcements. IBM has not tested those products and cannot confirm the performance, compatibility, or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products.

Prices are suggested US list prices and are subject to change without notice. Starting price may not include a hard drive, operating system or other features. Contact your IBM representative or Business Partner for the most current pricing in your geography. Any proposed use of claims in this presentation outside of the United States must be reviewed by local IBM country counsel prior to such use. The information could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein; these changes will be incorporated in new editions of the publication. IBM may make improvements and/or changes in the product(s) and/or the program(s) described in this publication at any 2 © 2015 IBM Corporation IBM

Notice Regarding Specialty Engines Any information contained in this document regarding Specialty Engines (“SEs”) and SE eligible workloads provides only general descriptions of the types and portions of workloads that are eligible for execution on Specialty Engines (e.g., zIIPs, zAAPs, and IFLs). IBM authorizes customers to use IBM SE only to execute the processing of Eligible Workloads of specific Programs expressly authorized by IBM as specified in the “Authorized Use Table for IBM Machines” provided at www.ibm.com/systems/support/machine_warranties/machine_code/aut.html (“AUT”). No other workload processing is authorized for execution on an SE. IBM offers SEs at a lower price than General Processors/Central Processors because customers are authorized to use SEs only to process certain types and/or amounts of workloads as specified by IBM in the AUT.

Linux on IBM z Systems introduction

Interesting facts and numbers

Fun facts around Linux on IBM z Systems

. Linux kernel v3.19 has 19,129,942 lines of code How many lines of code had the initial patch for s390 ? 36,811 lines of code (patch size) . How many lines of code are specific to s390 in v3.19 ? 213,285 lines of code (1.11%) . How many lines of assembler code are in v3.19 ? 396,473 lines in 1376 files, 4653 specific to s390 (1.17%) . How many individual developers contributes to Linux in 2014 ? 3697 individuals, 1225 with a single commit, 32 contributed directly to s390 . What is the architecture with the larges amount of core changes ? ARM, with ~105 KLOC per release for v3.x, followed by mips ~24 KLOC, powerpc ~23 KLOC, and x86 ~22 KLOC. s390 has an average of 7 KLOC.

Linux is Linux, but …features, properties and quality differ dependent on your platform and your use case Source: http://kernel.org http://linuxcounter.net

git commits per architecture in 3.x

20000

18000

16000

14000

12000

10000

8000

6000

4000

2000

0 r v e a 2 n x s n 4 k s 0 2 c c c 0 e 2 6 a c 4 g e h c r l i i 2 r r m r f f i h 6 z 0 s 8 6 3 r o 6 8 a p s s p 9 s 3 s r i i i t 3 a t r o a v k p c c g a 6 a 3 o r r 3 e x n a r l l i i e r m e c p c a a m 0 n s e m r a b a n m s s o t a w x 1 e p a m l o c x o r e n p i b p h c o n i m u m

. 27.3% of Total installed MIPS run Linux as of 4Q14 Installed Capacity Over Time . Installed IFL MIPS increased 12% from 4Q13 to 4Q14 . 39% of z Systems clients have IFL’s installed as of 4Q14 . 82 of the top 100 z Systems clients are running Linux on the mainframe as of 4Q14 ** . 35% of all z Systems servers have IFLs . 60% of new FIE/FIC z Systems accounts run Linux

Linux on IBM z Systems distributions

What is available today

Linux on IBN z Systems distributions in service

. SUSE Linux Enterprise Server 10 (GA 07/2006) – Kernel 2.6.16, GCC 4.1.0, Service Pack 4 (GA 04/2011) . SUSE Linux Enterprise Server 11 (GA 03/2009) – Kernel 2.6.27, GCC 4.3.3, Service Pack 1 (GA 06/2010), Kernel 2.6.32 – Kernel 3.0, GCC 4.3.4, Service Pack 3 (GA 07/2013) . SUSE Linux Enterprise Server 12 (GA 10/2014) – Kernel 3.12, GCC 4.8 . Red Hat Enterprise Linux AS 5 (GA 03/2007) – Kernel 2.6.18, GCC 4.1.0, Update 10 (GA 10/2013) . Red Hat Enterprise Linux AS 6 (GA 11/2010) – Kernel 2.6.32, GCC 4.4.0 Update 5 (GA 11/2013) . Red Hat Enterprise Linux AS 7 (GA 06/2014) – Kernel 3.10, GCC 4.8 . Others – Debian, Slackware, – Support may be available by some third party 9 © 2015 IBM Corporation IBM

Supported Linux Distributions

Current Linux on IBM z Systems Technology

Key features & functionality already contained in the SUSE & Red Hat Distributions

IBM zEnterprise EC12 and BC12 support

. Transactional execution (kernel 3.7) 6.4 11.3 – Also known as hardware transactional memory – CPU features that allows to execute a group of instructions atomically – Optimistic execution, if a transaction conflicts a rollback to a saved state is done

. Storage class memory – Flash Express (kernel 3.7) 6.4 11.3 – Internal Flash Solid State Disk (SSD) – Accessed via Extended Asynchronous Data Mover (EADM) sub-channels – Support for concurrent MCL updates with kernel version 3.8 . Support for Crypto Express 4S cards (kernel 3.7) 6.4 11.3 – New generation of crypto adapters plug-able into the I/O drawer – New type 10 which uses a bit field to indicate capabilities of the crypto card . Native PCI feature cards (base in kernel 3.8, ongoing) 7.0 12 – Support for native PCIe adapters visible to the operating system . Oprofile zEC12 hardware sampling support (kernel 3.10) 7.0 – Extend the hardware sampling to support zEC12

System zEC12 features – Transactional Execution 11.3

6.4 . Transactional execution is a concurrency mechanism of the CPU comparable to database transactions – Several reads and stores from/to memory logically occur at the same time – Improves performance for fine-grained serialization – Useful for lock-less data structures and speculative compiler optimizations . Two types of transactions: constraint and non-constraint . Conflicting memory accesses will cause the transaction to abort – Transaction abort is rather expensive – Constraint transaction will automatically restart – Ratio of successful vs. aborted transaction is important for performance . Kernel support is required to enable user programs to use transactional execution – Control registers setup – Debugging support for additional PER controls via ptrace

System zEC12 features – Transactional Execution 11.3

6.4 . Example of a list_add operation struct spinlock_t list_lock; struct list_head list_head; void list_add(struct list_head *new) { spin_lock(&list_lock, 0, 1); list_add(new, &list_head); Typical pattern: spin_unlock(&list_lock, 1, 0); 1) lock, 2) a short operation, 3) unlock } Traditional code: Transactional code # spin_lock # begin transaction larl %r3,list_lock tbeginc 0,0 lhi %r1,1 lock: lhi %r0,0 cs %r0,%r1,0(%r3) ltr %r0,%r0 jne lock # list_add # list_add larl %r4,list_head larl %r4,list_head lg %r5,0(%r4) lg %r5,0(%r4) stg %r4,0(%r2) stg %r4,0(%r2) stg %r5,8(%r2) stg %r5,8(%r2) stg %r2,0(%r5) stg %r2,0(%r5) stg %r2,8(%r4) stg %r2,8(%r4) # spin_unlock # end transaction cs %r1,%r0,0(%r3) tend br %r14 br %r14

6.4 . PCIe I/O adapter with NAND Flash SSDs – Flash Express cards are plugged as pairs to build a RAID10 • Pair is connected with interconnect cables • Card replacement is concurrent if one card fails – Up to 4 pairs of cards are supported (4 * 1.4TB = 5.6TB) . New tier of memory: Storage Class Memory – Accessed via Extended Asynchronous Data Mover (EADM) subchannels via the new Storage Class Memory (SCM) block driver . Flash Express is split into memory increments – Memory increments are assigned to LPARs via the SE or HMC – Memory increment size is 16 GB . Flash Express is not persistent over IML

PCI support

. Native PCIe feature cards introduced on zEC12 and zBC12 – 10GbE RoCE Express, network card for SMC-R – zEDC Express, data compression/decompression card

. Native PCIe adapter concept – Plugged into an PCIe I/O drawer – Managed by an internal firmware processor (IFP) – Device driver for the PCIe function is located in the operating system

. Uses standard Linux PCI support and drivers with some constraints – Only MSIX, no port I/O, memory mapped I/O by use of PCI load/store instructions – Provides ability to assign individual functions of an adapter to an LPAR – Converted z Systems architecture code to use generic hardirqs – Only selected PCIe adapters are known to the IFP and surfaced to the OS

10GbE RoCE Express 7.0 12

. Native PCIe networking card – 10 Gigabit remote direct memory access (RDMA) capable network card – Uses Infiniband RDMA over Converged Ethernet (RoCE) specification – Up to 16 10GbE RoCE Express adapters per machine – Reduced latency and lower CPU overhead – Supports point-to-point connections and switch connection with an enterprise-class 10 GbE switch

. Software support – z/OS V2R1 with PTFs supports SMC-R with RoCE – z/VM support planned – Linux support in available upstream and as tech preview in SLES12 / RHEL7

. Native PCIe data compression / decompression card – Up to 8 adapters can be installed into a single machine – With large blocks, it can compress data at more than 1 GB per second – Implements compression as defined by RFC1951 (DEFLATE) – Comparable to “gzip -1”

. Software support – z/OS V2R1, V1R13 and V1R12 with PTFs – Linux device driver to gain access to zEDC has been posted on LKML and has been accepted into the upstream kernel – The zlib open source library is a C implementation commonly used to provide compression and decompression services.

Linux on IBM z Systems features – Compiler toolchain

. zEnterprise 196 exploitation (gcc 4.6) 6.1 11.2 – Use option -march=z196 to utilize the new instructions added with z196 – Use -mtune=z196 to schedule the instruction appropriate for the new out-of-order pipeline of z196 – Re-compiled code/apps get further performance gains through 110+ new instructions

. zEC12/zBC12 exploitation CPU (gcc 4.8) 7.0 – Use option -march=zEC12 to utilize the instructions added with zEC12 – Use option -mtune=zEC12 to schedule the instructions appropriate for the pipeline of zEC12 – zEC12/zBC12 comes with new instructions • Transactional Memory support • Improved branch instructions

Emerging Linux on IBM z Systems Technology

Upcoming features & functionality from IBM

GPFS for Linux on IBM z Systems

. The GPFS file system running on Linux on IBM z Systems – Enable software solutions dependent on GPFS – Position GPFS as alternative to OCFS2 for SLES – Enable Linux on System for High Availability (HA) solutions with GPFS . Elastic Storage for Linux on IBM z Systems – Express Edition of Elastic Storage 4.1 will be the base for Linux on IBM z Systems support – Linux instances in LPAR mode or on z/VM (on the same of different CECs) – Up to 16 cluster nodes (same or mixed Linux distributions/releases) – Support for ECKD™-based storage and FCP-based storage • Only DS8000® is supported . Minimum supported Linux distributions – SuSE Linux Enterprise Server (SLES) 10 SP3 + Maintweb-Update – Red Hat Enterprise Linux (RHEL) 6.5 + Errate Update

XL C/C++ for Linux on IBM z Systems Managed Beta Program

. XL C/C++ for Linux on IBM z Systems – Will be part of a family of advanced C/C++ compiler products already available on z/OS, AIX, and Linux on Power. – Expected to ease application migration to Linux on IBM z Systems through: – Conformance to the latest C and C++ programming standards – Compatibility with GNU C/C++ – Will maximize application performance through IBM’s industry-leading optimization technology

. Benefits of participating in this Beta include: – Opportunity to influence the product and future product direction – Ability to test code and documentation, and help ensure compatibility in their environment – Free education, code, and documentation during the beta – Free support by development during the beta for questions and problems

For more information and how to submit a nomination to participate see http://bit.ly/xlbeta

Future Linux on IBM z System Technology

Software which has already been developed and integrated into the upstream packages - but is not yet available in any Enterprise Linux Distribution

IBM z13 support

. Vector extension facility (kernel 3.18) – 32 128-bit vector registers are added to the CPU – 139 new instructions to operate on the vector registers – User space programs can use vectors to speed up all kinds of functions, e.g. string functions, crc checksums, ... . CPU multi threading support (> kernel 3.19) – Once enabled the multi threading facility provides multiple CPUs for a single core. – The CPUs of a core share certain hardware resource such as execution units or caches – Avoid idle hardware resources, e.g. while waiting for memory . Extended number of AP domains (kernel 3.18) – AP crypto domains in the range 0-255 will be detected . Crypto Express 5S cards (> kernel 3.19) – New generation of crypto adapters with improved performance

IBM z13 – Vector extension

. 32 vector registers with 128 bits each – Register can be split into 16 bytes, 8 shorts, 4 integers, or 2 long integers – Up to 4 concurrent 32x32 multiply / adds – Many new vector instructions with many specialized use cases • e.g. Vector Galois Field Multiply Sum and Accumulate (VGFMA)

. Vector registers and floating pointer register partially overlap – The program can use either the FPR or the VR with 0-15 at the same time

0 %f0 or %v0 bits 0:63 %v0 bits 64:127 ...... 15 %f0 or %v0 bits 0:63 %v0 bits 64:127 16 %v0 bits 0:127 ...... 31 %v0 bits 0:127

IBM z13 – Vector extension

. Vector instruction example: size_t strlen(const char *s)

# R2 address of the string, R1 will contain length XGR R1,R1 Zero out running index LOOP: VLBB V16,0(R1,R2),6 Load up to 16 bytes LCBB R3,0(R1,R2),6 Find how many bytes were loaded ALGRK R1,R1,R3 Increment length by bytes loaded VFENEBZ V17,V16,V16 Look for 0 byte VLVGB R4,V17,7(R0) Extract index to gpr (16-no match) CLGR R3, R4 If GLEN <= GPOS have more to search BRNH LOOP SLGRK R1,R1,R3 Subtract off amount loaded ALGRK R1,R1,R4 Add amount to the zero that was found

– 7 instructions to scan 16 bytes in the body of the strlen loop

. Standard implementation uses loop unrolling for 8 bytes – 1x 'CLI' + 1x 'BRC' for each byte, 1x 'LA' + 1x 'BRCT' for 8 bytes → 34 instructions to scan 16 bytes in the body of the strlen loop

IBM z13 – CPU multi threading (MT) support

. Up to two hardware threads on a single core – Also known as simultaneous multithreading (SMT) on POWER or Hyper-threading on x86 . The operating system needs to opt-in to enable SMT – The LPAR code starts a partition with one logical CPU per core – After the SIGP to enable MT additional logical CPUs are surfaced and the CPU addressing changes

CPU address (multi-threading not enabled)

CPU address (multi-threading enabled) Core Identification TID

– The additional CPUs work just like “normal” CPUs, but with different performance characteristics

IBM z13 – CPU multi threading (MT) support

. Example lscpu output with MT enabled

# lscpu -e CPU BOOK SOCKET CORE L1d:L1i:L2d:L2i ONLINE CONFIGURED POLARIZATION ADDRESS 0 0 0 0 0:0:0:0 yes yes horizontal 0 1 0 0 0 1:1:1:1 yes yes horizontal 1 2 0 1 1 2:2:2:2 yes yes horizontal 2 3 0 1 1 3:3:3:3 yes yes horizontal 3 4 0 1 2 4:4:4:4 yes yes horizontal 4 5 0 1 2 5:5:5:5 yes yes horizontal 5 6 0 1 3 6:6:6:6 yes yes horizontal 6 7 0 1 3 7:7:7:7 yes yes horizontal 7 8 0 1 4 8:8:8:8 yes yes horizontal 8 9 0 1 4 9:9:9:9 yes yes horizontal 9

IBM z13 – CPU multi threading (MT) support

. Scheduler CPU domains non-MT vs. MT

Domain Domain CPU CPU

Domain Domain Domain Domain BOOK BOOK BOOK BOOK

New domain Domain Domain Domain Domain Domain Domain MC MC MC MC MC MC

RQ RQ RQ RQ RQ

Runqueue Domain Domain Domain Domain Domain per logical CPU SMT SMT SMT SMT SMT

RQ RQ RQ RQ RQ RQ RQ RQ RQ RQ

Two independent runqueues per Core

IBM z13 – CPU multi threading (MT) support

. CPU time accounting with MT enabled – New CPU-MF counters: number of cycles with n threads active – Average thread density is calculated once per tick – The raw CPU timer deltas are scaled with the average thread density – The scaled CPU time is available via the 'taskstats' interface – The standard CPU time values in /proc use the unscaled CPU time

IBM z13 – CPU multi threading (MT) support

. Workloads may or may not benefit from MT – Certain processor resource need to be underutilized for MT to be effective

. Workloads characteristics with a positive effect on MT – A large number of cache misses – Long chains of instruction dependencies – A large number of branch mis-predictions . Workload characteristics with a detrimental effect on MT – A homogeneous instruction mix targeting a scarce function unit – Intensive use of the memory system with poor cache locality – Applications with poor scaling, in general MT requires more logical CPUs

IBM z Systems kernel features – memory management

. Implement fault based referenced page detection (kernel 3.12) – Convert referenced page detection from the reference-bit in the storage key to a fault based method. An old page is now always mapped with the invalid bit set (no read, no write access). – New mappings are always created with the software referenced bit set – Removes the storage key operations to detect page referenced state.

. Dirty bits for large segment table entries (kernel 3.17) – Allows to track the changed state of 1MB large pages – Useful to avoid paging if a transparent-huge-page is split into 4K pages

. Avoiding storage key operations improves performance – The savings in storage key operations outweigh the slightly increase number of faults – After IPL a system without KVM will not access the storage keys at all – KVM still makes use of storage keys for provide correct guest virtualization

IBM z Systems kernel features – memory management

. Convert bootmem to memblock (kernel 3.16) – The bootmem allocator needs one bits for each page during IPL – With large memory size the bootmem bit arrays get huge, they are freed after IPL – memblock uses lists of memory ranges which stay persistent

. Uaccess rework (kernel 3.16) – The old uaccess code used a page table walk for some operations, e.g. the atomic futex operations – The new solution is to temporarily switch from home space to primary space in the uaccess kernel code – Avoid the costly page table walk and uses hardware DAT translation

IBM z Systems kernel features – core improvements

. CPU-Measurement Sampling Facility (kernel 3.14) – Uses the hardware CPU sampling facility to take snapshots of a set of sample data at a specified sampling interval, e.g. the cycle counter – Integrated into the Linux 'perf' tool – The basic-sampling mode and the diagnostic-sampling mode are supported – The diagnostic-sampling mode is intended for use by IBM support only . Example how to record sampling data for an application # perf record -e rB0000 ± /bin/df Filesystem 1Kblocks Used Available Use% Mounted on /dev/dasda1 6967656 3360508 3230160 51% / none 942956 88 942868 1% /dev/shm /dev/dasdb1 6967656 4132924 2474128 63% /root [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.001 MB perf.data (~29 samples) ] . Display the collected sample data # perf report

IBM z Systems kernel features – core improvements

. Add interface for partition-resource management (kernel 3.14) – The diagnose 0x304 interface is used to inspect and change the different LPAR partition- resource parameters – The LPAR needs to be authorized to participate in CPU management – The binary kernel interface allows a system management software to control the partition weight and partition-capping flags . Increase number of CPUs and limit memory use (kernel 3.15) – zEC12 can have up to 101 CPUs in an LPAR, z13 up to 141 * 2 = 282 (with SMT) . Spinlock refactoring (kernel 3.16) – Do a load-and-test first, before trying to get the lock with compare-and-swap, this is required for the SMT support with z13 – Improve code generation of the inline spinlock code . Add diagnose 0x288 watchdog support for LPAR (kernel 3.16) – The diagnose 0x288 watchdog driver has been converted to the generic watchdog API – The LPAR variant of the diagnose 0x288 watchdog has been added

IBM z Systems kernel features – core improvements

. Feed randomness with system information (kernel 3.17) – Use the virtual-machine CPU information data block and the CPU-ID of the boot CPU as a source of device randomness. – Helps with KVM to provide a different initial random pool for each virtual machine . Architecture backend for uprobes (kernel 3.18) – Uprobes implements a mechanism by which a kernel function can be invoked whenever a process executes a specific instruction location – E.g. trace all callers of the user space libc function 'malloc' . Improved ftrace/kprobes architecture backend (kernel 3.19) – Switching on CONFIG_FTRACE has a performance impact – Rework the function prolog for ftrace/kprobes to reduce the cost – With the new code ftrace/kprobes can be used in a production kernel . HMC drive CD/DVD access (kernel 3.18) – The HMC device driver allows user space access to a HMC media drive – Can be used in a z/VM and an LPAR environment

IBM z Systems kernel features – I/O improvements

. Lazy IOTLB flushing for PCI DMA unmap (kernel 3.17) – Delay the costly IOTLB flush operation by keeping track of DMA addresses after unmap – Only after the DMA addresses have wrapped the IOTLB is flushed . DASD support for control unit initiated reconfiguration (kernel 3.18) – A storage server can indicate concurrent hardware changes to the OS – Automates channel path management for selected DS8000 service actions . Auto port scan resiliency for zfcp (kernel 3.19) – Improves the Fibre Channel port scan behaviour – Avoids concurrent port scans in a virtual server environment with a short delay – Rate limit the number of ports scans . Handle multiple SCM requests in one HW requests (kernel 3.19) – Do up to 8 block layer requests per HW request to improve performance . PCI memory access system calls (kernel 3.19) – Emulate PCI memory access from user space by using system calls

IBM z Systems kernel features – networking & security

. HiperSockets layer 2 bridge port functionality (kernel 3.14) – With Linux acting as a software network bridge the network port acting as the bridge needs to be able to receive frames addressed to unknown MAC addresses – HiperSocket devices can be configured as primary and secondary bridge ports . qeth: priority queueing options + IPv6 support (kernel 3.16) – Adds new options for priority queuing • SKB priority queue • VLAN priority queuing based on the data in the 802.1Q header . Add support for EP11 coprocessor cards (kernel 3.14) – Extend the zcrypt driver with a new capability to service EP11 requests for the Crypto Express4S card in EP11 (Enterprise PKCS#11 mode) coprocessor mode – For more information about EP11, see “Exploiting Enterprise PKCS #11 using OpenCryptoki”, SC34-2713 . Support for extended number of AP domains (kernel 3.18) – Extends the number of AP domains recognized by the zcrypt device driver to 256

IBM z Systems kernel removals

. 31-bit kernel support will be removed with kernel 3.21 – 31-bit application will continue to work, only the 31-bit kernel is discontinued – A 64-bit kernel can be used in combination with a 100% 31-bit user space

. The claw network driver will be remove with kernel 3.21 – Common Link Access for Workstation, used some old RS/6000s, Cisco Routers (CIP) and 3172 devices – We could not find a single user of this device driver

New - ibm.com/support/knowledgecenter/

Linux on IBM z Systems info at: ibm.com/support/knowledgecenter/linuxonibm/liaaf/lnz_r_main.html

Documentation news – Updates available . Linux Distributions – SUSE Linux Enterprise Server 11 SP 3 – Red Hat Enterprise Linux 6.4

. Upstream Linux 3.14, 3.16, 3.17 – ibm.com/developerworks/linux/linux390/documentation_dev.html

. Security – Secure Key Common Cryptographic Architecture 4.2.10 Application Programmer's Guide – Libica 2.2.0. Programmer's Guide

. Redbooks – IBM Wave, Virtualization, Oracle, Security

. Whitepapers – FileNet P8 5.1, Live Guest Relocation, iSCSI

. s390-tools is a package with a set of user space utilities to be used with the Linux on IBM z Systems distributions. – It is the essential tool chain for Linux on IBM z Systems – It contains everything from the boot loader to dump related tools for a system crash analysis .

. This software package is contained in all major (and IBM supported) enterprise Linux distributions which support s390 – RedHat Enterprise Linux version 4, 5, 6, and 7 – SuSE Linux Enterprise Server version 9, 10, and 11

. Website: http://www.ibm.com/developerworks/linux/linux390/s390-tools.html

. Feedback: [email protected]

chccwdev dasdfmt dbginfo chchp dasdinfo dumpconf chreipl dasdstat zfcpdump DUMP chshut dasdview zfcpdbf & chcrypt fdasd zgetdump DEBUG chmem CHANGE tunedasd DASD scsi_logging_level lscss lschp mon_fsstatd lsdasd mon_procd vmconvert ziomon vmcp lshmc MONITOR lsluns hyptop vmur cms-fuse lsqeth ip_watcher z/VM lsreipl osasnmpd cpuplugd lsshut qetharp lstape iucvconn qethconf iucvtty lszcrypt qethqoat NETWORK lszfcp ts-shell DISPLAY ttyrun MISC lsmem tape390_display hmcdrvfs tape390_crypt zipl zdsfs FILESYSTEM TAPE BOOT

. zdsfs – mount a z/OS DASD – The zdsfs file systems is based on FUSE – Use the zdsfs command to mount a z/OS DASD as read-only Linux file system. – The zdsfs file system translates the z/OS data sets into Linux semantics – Allows Linux to read z/OS physical sequential data sets and partitioned data sets • Physical sequential data sets are represented as files • Partitioned data sets are represented as directories that contain the PDS members as files – Other z/OS data set formats are not supported – To access a z/OS DASD the raw-track access mode of the DASD must be enabled

. hmcdrvfs – mount the HMC media drive as virtual filesystem – The hmcdrvfs file systems is based on FUSE – On the HMC, the media must be assigned to the associated system image – Use the hmcdrvfs command to mount the HMC media drive as read-only Linux file system – Allows Linux to read files from the HMC media drive – Required privileges • For z/VM the guest virual machine must have at least privilege class B and the media must be assigned to the LPAR where z/VM runs • For LPAR, the activation profile must allow issuing SCLP requests. – One use case would be installation of a distribution from the HMC DVD drive . lshmc – list media contents in the HMC media drive – Useful to quickly check what kind of HMC media drive is attached

Kernel news – Common code

. Linux version 3.14 (2014-03-30) – Deadline scheduling class for better real-time scheduling – zram memory compression mechanism considered stable – Btrfs inode properties – Userspace locking validator – TCP automatic corking

. Linux version 3.15 (2014-05-08) – Improved working set size detection – New file locking scheme: open file description locks – Faster erasing and zeroing of parts of a file – File cross-renaming support – FUSE improved write performance

Kernel news – Common code

. Linux version 3.16 (2014-08-03) – Unified Control Group hierarchy – XFS free inode btree, for faster inode allocation – TCP Fast Open server mode on IPv6 support – seccomp: use internal BPF JIT to speed up filters – btrfs improvements

. Linux version 3.17 (2014-10-05) – File sealing with fcntl() to eases handling of shared memory – Secure generation of random numbers with the getrandom() syscall – Support for page fault tracing in perf trace – Allow seccomp to set a filter across all threads

Kernel news – Common code

. Linux version 3.18 (2014-12-07) – Overlayfs to combine two filesystems into a single file system namespace – bpf() syscall for eBFP virtual machine programs – TCP: Data Center TCP congestion algorithm – Networking performance optimization: transmission queue batching

. Linux version 3.19-rc7 (2015-02-01) – Btrfs: support scrubbing and fast device replacement in RAID 5 and 6 – New architecture: Altera Nios II processors – Networking: support for routing and switching offloading – NFSv4.2 support for hole punching and preallocation

Questions?

Martin Schwidefsky Schönaicher Strasse 220 71032 Böblingen, Germany

Linux on IBM z Systems Phone +49 (0)7031-16-2247 Development [email protected]